Automatic three-dimensional reconstruction based on four-view stereo vision using checkerboard pattern

An automatic three-dimensional(3D) reconstruction method based on four-view stereo vision using checkerboard pattern is presented. Mismatches easily exist in traditional binocular stereo matching due to the repeatable or similar features of binocular images. In order to reduce the probability of mismatching and improve the measure precision, a four-camera measurement system which can add extra matching constraints and offer multiple measurements is applied in this work. Moreover, a series of different checkerboard patterns are projected onto the object to obtain dense feature points and remove mismatched points. Finally, the 3D model is generated by performing Delaunay triangulation and texture mapping on the point cloud obtained by four-view matching. This method was tested on the 3D reconstruction of a terracotta soldier sculpture and the Buddhas in the Mogao Grottoes. Their point clouds without mismatched points were obtained and less processing time was consumed in most cases relative to binocular matching. These good reconstructed models show the effectiveness of the method.

基于Stereo Camera-YOLOv5自然环境下百香果检测与定位模型

针对百香果采摘机器人在自然环境中作业时受复杂光线及遮挡影响,难以快速精确地检测及定位成熟百香果的问题,提出一种基于Stereo Camera-YOLOv5自然环境下成熟百香果检测及定位模型。针对自然环境下光线以及遮挡的影响,通过MSRCP算法、随机遮挡、数据增扩等图像处理算法对原始数据集进行优化。将优化的数据集输入到YOLOv5网络中训练出最优模型,在检测代码中嵌入双目立体视觉算法。该模型对自然环境下百香果进行检测及成熟度判断,将判断为成熟的百香果进行图像处理,并提取到中心点二维坐标。通过立体匹配及视差计算得到中心点的三维坐标。田间试验结果表明,该模型的目标检测准确率为97.8%,总体准确率为90.2%,平均运行时间为4.85 s。该系统鲁棒性强、实时性好,能够更好地实现自然环境下成熟百香果的检测及定位,为百香果采摘机器人后续工作奠定基础。

Hawk‐eye‐inspired perception algorithm of stereo vision for obtaining orchard 3D point cloud navigation map

The binocular stereo vision is the lowest cost sensor for obtaining 3D information.Considering the weakness of long‐distance measurement and stability,the improvement of accuracy and stability of stereo vision is urgently required for application of precision agriculture.To address the challenges of stereo vision long‐distance measurement and stable perception without hardware upgrade,inspired by hawk eyes,higher resolution perception and the adaptive HDR(High Dynamic Range)were introduced in this paper.Simulating the function from physiological structure of‘deep fovea’and‘shallow fovea’of hawk eye,the higher resolution reconstruction method in this paper was aimed at ac-curacy improving.Inspired by adjustment of pupils,the adaptive HDR method was proposed for high dynamic range optimisation and stable perception.In various light conditions,compared with default stereo vision,the accuracy of proposed algorithm was improved by 28.0%evaluated by error ratio,and the stability was improved by 26.56%by disparity accuracy.For fixed distance measurement,the maximum improvement was 78.6%by standard deviation.Based on the hawk‐eye‐inspired perception algorithm,the point cloud of orchard was improved both in quality and quantity.The hawk‐eye‐inspired perception algorithm contributed great advance in binocular 3D point cloud recon-struction in orchard navigation map.

6-DOF MOTION AND CENTER OF ROTATION ESTIMATION BASED ON STEREO VISION

A new motion model and estimation algorithm is proposed to compute the general rigid motion object＇s 6-DOF motion parameters and center of rotation based on stereo vision. The object＇s 6-DOF motion model is designed from the rigid object＇s motion character under the two defined reference frames. According to the rigid object＇s motion model and motion dynamics knowledge, the corresponding motion algorithm to compute the 6-DOF motion parameters is worked out. By the rigid object pure rotation motion model and space sphere geometry knowledge, the center of rotation may be calculated after eliminating the translation motion out of the 6-DOF motion. The motion equations are educed based on the motion model and the closed-form solutions are figured out. To heighten the motion estimation algorithm＇s robust, RANSAC algorithm is applied to delete the outliers. Simulation and real experiments are conducted and the experiment results are analyzed. The results prove the motion model＇s correction and algorithm＇s validity.

Accuracy study of a binocular-stereo-vision-based navigation robot for minimally invasive interventional procedures

BACKGROUND Medical robot is a promising surgical tool,but no specific one has been designed for interventional treatment of chronic pain.We developed a computed tomography-image based navigation robot using a new registration method with binocular vision.This kind of robot is appropriate for minimal invasive interventional procedures and easy to operate.The feasibility,accuracy and stability of this new robot need to be tested.AIM To assess quantitatively the feasibility,accuracy and stability of the binocularstereo-vision-based navigation robot for minimally invasive interventional procedures.METHODS A box model was designed for assessing the accuracy for targets at different distances.Nine(three sets)lead spheres were embedded in the model as puncture goals.The entry-to-target distances were set 50 mm(short-distance),100 mm(medium-distance)and 150 mm(long-distance).Puncture procedure was repeated three times for each goal.The Euclidian error of each puncture was calculated and statistically analyzed.Three head phantoms were used to explore the clinical feasibility and stability.Three independent operators conducted foramen ovale placement on head phantoms(both sides)by freehand or under the guidance of robot(18 punctures with each method).The operation time,adjustment time and one-time success rate were recorded,and the two guidancemethods were compared.RESULTS On the box model,the mean puncture errors of navigation robot were 1.7±0.9 mm for the short-distance target,2.4±1.0 mm for the moderate target and 4.4±1.4 mm for the long-distance target.On the head phantom,no obvious differences in operation time and adjustment time were found among the three performers(P>0.05).The median adjustment time was significantly less under the guidance of the robot than under free hand.The one-time success rate was significantly higher with the robot(P<0.05).There was no obvious difference in operation time between the two methods(P>0.05).CONCLUSION In the laboratory environment,accuracy of binocular-stereo-vision-based navigation robot is acceptable for target at 100 mm depth or less.Compared with freehand,foramen ovale placement accuracy can be improved with robot guidance.

Robot stereo vision calibration method with genetic algorithm and particle swarm optimization

Accurate stereo vision calibration is a preliminary step towards high-precision visual posi- tioning of robot. Combining with the characteristics of genetic algorithm （GA） and particle swarm optimization （PSO）, a three-stage calibration method based on hybrid intelligent optimization is pro- posed for nonlinear camera models in this paper. The motivation is to improve the accuracy of the calibration process. In this approach, the stereo vision calibration is considered as an optimization problem that can be solved by the GA and PSO. The initial linear values can be obtained in the frost stage. Then in the second stage, two cameras＇ parameters are optimized separately. Finally, the in- tegrated optimized calibration of two models is obtained in the third stage. Direct linear transforma- tion （DLT）, GA and PSO are individually used in three stages. It is shown that the results of every stage can correctly find near-optimal solution and it can be used to initialize the next stage. Simula- tion analysis and actual experimental results indicate that this calibration method works more accu- rate and robust in noisy environment compared with traditional calibration methods. The proposed method can fulfill the requirements of robot sophisticated visual operation.

Cast shadow detection for moving objects based on binocular stereo vision

A novel cast shadow detection approach was proposed.A stereo vision system was used to capture images instead of traditional single camera.It was based on an assumption that cast shadows were on a special plane.The image obtained from one camera was inversely projected to the plane and then transformed to the view from another camera.The points on the plane shared the same position between original image and the transformed image.As a result,the cast shadows can be detected.In order to improve the efficiency of cast shadow detection and decrease computational complexity,the obvious object areas in CIELAB color space were removed and the potential shadow areas were obtained.Experimental results demonstrate that the proposed approach can detect cast shadows accurately even under various illuminations.

Control of a 6DOF Mobile Manipulator with Object Detection and Tracking Using Stereo Vision

A supportive mobile robot for assisting the elderly is an emerging requirement mainly in countries like Japan where population ageing become relevant in near future.Falls related injuries are considered as a critical issue when taking into account the physical health of older people.A personal assistive robot with the capability of picking up and carrying objects for long/short distances can be used to overcome or lessen this problem.Here,we design and introduce a 3 D dynamic simulation of such an assistive robot to perform pick and place of objects through visual recognition.The robot consists of two major components;a robotic arm or manipulator to do the pick and place,and an omnidirectional wheeled robotic platform to support mobility.Both components are designed and operated according to their kinematics and dynamics and the controllers are integrated for the combined performance.The objective was to improve the accuracy of the robot at a considerably high speed.Designed mobile manipulator has been successfully tested and simulated with a stereo vision system to perform object recognition and tracking in a virtual environment resembling aroom of an elderly care.The tracking accuracy of the mobile manipulator at an average speed of 0.5 m/s is 90%and is well suited for the proposed application.

Research on Extraction of Bottom of Shoe Pattern Based on Binocular Stereo Vision

In order to quickly and efficiently get the information of the bottom of the shoe pattern and spraying trajectory, the paper proposes a method based on binocular stereo vision. After acquiring target image, edge detection based on the canny algorithm, the paper begins stereo matching based on area and characteristics of algorithm. To eliminate false matching points, the paper uses the principle of polar geometry in computer vision. For the purpose of gaining the 3D point cloud of spraying curve, the paper adopts the principle of binocular stereo vision 3D measurement, and then carries on cubic spline curve fitting. By HALCON image processing software programming, it proves the feasibility and effectiveness of the method

Development of a Stereo Vision-based Pick and Place System for Robotic Manipulators

Conventional robotic manipulators consist of touch and vision sensors in order to pick and place differently shaped objects.Due to the technology development and degrading sensors over a long period,the stereo vision technique has become a promising alternative.In this study,a low-cost stereo vision-based system,and a gripper to be placed at the end of the robot arm(Fanuc M10 iA/12)are developed for position and orientation estimation of robotic manipulators to pick and place different shaped objects.The stereo vision system developed in this research is used to estimate the position(X,Y,Z),orientation(P_(y))of the Center of Volume of four standard objects(cube,cuboid,cylinder,and sphere)whereas the robot arm with the gripper is used to mechanically pick and place the objects.The stereo vision system is placed on the movable robot arm,and it consists of two cameras to capture two 2D views of a stationary object to derive 3D depth information in 3D space.Moreover,a graphical user interface is developed to train a linear regression model,live predict the coordinates of the objects,and check the accuracy of the predicted data.The graphical user interface can also send predicted coordinates and angles to the gripper and the robot arm.The project is facilitated with python programming language modules and image processing techniques.Identification of the stationary object and estimation of its coordinates is done using image processing techniques.The final product can be identified as a device that converts conventional robot arms without an image processing vision system into a highly precise and accurate robot arm with an image processing vision system.Experimental studies are performed to test the efficiency and effectiveness of used techniques and the gripper prototype.Necessary actions are taken to minimize the errors in position and orientation estimation.In addition,as a future implementation,an embedded system will be developed with a user-friendly software interface to install the vision system into the Fanuc M10 iA/12 robot arm and will upgrade the system to a device that can be implemented with any kind of customized robot arms available in the industry.

Fast and Robust Stereo Vision Algorithm for Obstacle Detection

Binocular computer vision is based on bionics, after the calibration through the camera head by double-exposure image synchronization, access to the calculation of two-dimensional image pixels of the three-dimensional depth information. In this paper, a fast and robust stereo vision algorithm is described to perform in-vehicle obstacles detection and characterization. The stereo algorithm which provides a suitable representation of the geometric content of the road scene is described, and an in-vehicle embedded system is presented. We present the way in which the algorithm is used, and then report experiments on real situations which show that our solution is accurate, reliable and efficient. In particular, both processes are fast, generic, robust to noise and bad conditions, and work even with partial occlusion.

Structured scene modeling using micro stereo vision system with large field of view

This paper presents a method for structured scene modeling using micro stereo vision system with large field of view. The proposed algorithm includes edge detection with Canny detector, line fitting with principle axis based approach, finding corresponding lines using feature based matching method, and 3D line depth computation.

Deforming analysis of sheet metal based on stereo vision and coordinate grid

A new approach based on stereo vision technology is introduced to analyzesheet metal deformation. By measuring the deformed circle grids that are printed on the sheetsurface before forming, the strain distribution of the workpiece is obtained. The measurement andanalysis results can be used to verify numerical simulation results and guide production. To getgood accuracy, some new techniques are employed: camera calibration based on genetic algorithm,feature abstraction based on self-adaptive technology, image matching based on structure feature andcamera modeling pre-constrains, and parameter calculation based on curve and surface optimization.The experimental values show that the approach proposed is rational and practical, which can providebetter measurement accuracy with less time than the conventional method.

Convergence of Stereo Vision-Based Multimodal YOLOs for FasterDetection of Potholes

Road potholes can cause serious social issues,such as unexpected damages to vehicles and traffic accidents.For efficient road management,technologies that quickly find potholes are required,and thus researches on such technologies have been conducted actively.The three-dimensional(3D)reconstruction method has relatively high accuracy and can be used in practice but it has limited application owing to its long data processing time and high sensor maintenance cost.The two-dimensional(2D)vision method has the advantage of inexpensive and easy application of sensor.Recently,although the 2D vision method using the convolutional neural network(CNN)has shown improved pothole detection performance and adaptability,large amount of data is required to sufficiently train the CNN.Therefore,we propose a method to improve the learning performance of CNN-based object detection model by artificially generating synthetic data similar to a pothole and enhancing the learning data.Additionally,to make the defective areas appear more contrasting,the transformed disparity map(TDM)was calculated using stereo-vision cameras,and the detection performance of the model was further improved through the late fusion with RGB(Red,Green,Blue)images.Consequently,through the convergence of multimodal You Only Look Once(YOLO)frameworks trained by RGB images and TDMs respectively,the detection performance was enhanced by 10.7%compared with that when using only RGB.Further,the superiority of the proposed method was confirmed by showing that the data processing speed was two times faster than the existing 3D reconstruction method.

Research on stereo vision path-planning algorithms for mobile robots autonomous navigation

Using stereo vision for autonomous mobile robot path-planning is a hot technology.The environment mapping and path-planning algorithms were introduced,and they were applied in the autonomous mobile robot experiment platform.Through experiments in the robot platform,the effectiveness of these algorithms was verified.

DEDUCTION OF CONSTRAINT CONDITIONS IN THE CALIBRATION OF STEREO-VISION SYSTEN

In the test-field calibration,multi-azimuth stereo image pairs areproduced of the outdoor large control-field by the stereo-vision system under cali-bration.While in the analytical processing,the relationship between image pairsis adopted as a constraint condition,which ensures the stability and quality of thecalibration results.This paper introduces the deduction process of the constraintconditions.

Intelligent Control of Mobile Robot with Redundant Manipulator & Stereovision: Quantum / Soft Computing Toolkit

The task of an intelligent control system design applying soft and quantum computational intelligence technologies discussed.An example of a control object as a mobile robot with redundant robotic manipulator and stereovision introduced.Design of robust knowledge bases is performed using a developed computational intelligence-quantum/soft computing toolkit(QC/SCOptKBTM).The knowledge base self-organization process of fuzzy homogeneous regulators through the application of end-to-end IT of quantum computing described.The coordination control between the mobile robot and redundant manipulator with stereovision based on soft computing described.The general design methodology of a generalizing control unit based on the physical laws of quantum computing(quantum information-thermodynamic trade-off of control quality distribution and knowledge base self-organization goal)is considered.The modernization of the pattern recognition system based on stereo vision technology presented.The effectiveness of the proposed methodology is demonstrated in comparison with the structures of control systems based on soft computing for unforeseen control situations with sensor system.The main objective of this article is to demonstrate the advantages of the approach based on quantum/soft computing.

基于多特征SAD-Census变换的立体匹配算法

视差不连续区域和重复纹理区域的误匹配率高一直是影响双目立体匹配测量精度的主要问题,为此,本文提出一种基于多特征融合的立体匹配算法。首先,在代价计算阶段,通过高斯加权法赋予邻域像素点的权值,从而优化绝对差之和(Sum of Absolute Differences,SAD)算法的计算精度。接着,基于Census变换改进二进制链码方式,将邻域内像素的平均灰度值与梯度图像的灰度均值相融合,进而建立左右图像对应点的判断依据并优化其编码长度。然后,构建基于十字交叉法与改进的引导滤波器相融合的聚合方法,从而实现视差值再分配,以降低误匹配率。最后,通过赢家通吃(Winner Take All,WTA)算法获取初始视差,并采用左右一致性检测方法及亚像素法提高匹配精度,从而获取最终的视差结果。实验结果表明,在Middlebury数据集的测试中,所提SAD-Census算法的平均非遮挡区域和全部区域的误匹配率为分别为2.67%和5.69%,测量200~900 mm距离的平均误差小于2%;而实际三维测量的最大误差为1.5%。实验结果检验了所提算法的有效性和可靠性。

电力变压器关键尺寸视觉检测方法及其缩比模型验证

电力变压器的关键尺寸测量是其装配制造、运输安装过程中重要一环,现有测量方法操作繁琐且效率较低。为此,本文提出了一种适用于110 kV油浸式电力变压器关键尺寸视觉检测方法,该方法利用YOLO目标检测算法、Grabcut图像分割算法实现关键组部件的智能检测与分割,然后基于双目立体视觉原理实现套管相间距离及变压器最大截面外形尺寸等关键尺寸的测量。本文搭建了110 kV油浸式电力变压器外观缩比模型,试验分析了拍摄距离和角度等因素对电力变压器关键尺寸视觉检测的影响。结果表明,本文基于缩比模型试验实现了在不同拍摄距离与角度下变压器关键尺寸的视觉检测,验证了该方法的有效性。本文方法可以为110 kV油浸式电力变压器现场尺寸测量提供参考。

一种密集多尺度特征引导代价聚合的改进立体匹配网络

针对目前立体匹配算法在重复纹理、无纹理、边缘等不适定性区域仍存在匹配不准确的问题,提出了一种基于PSMNet的密集多尺度特征引导代价聚合的立体匹配算法—DGNet(Dense multi-scale features Guided aggregation Network)。首先,基于密集连接空洞空间金字塔池化结构设计了密集多尺度特征提取模块,该模块利用不同膨胀率的空洞卷积提取不同尺度的区域级特征,并通过密集连接方式有效整合不同尺度的图像特征,使网络捕获丰富的上下文关系;其次,在每个视差等级下将左右特征图串联形成初始代价体,再提出密集多尺度特征引导代价聚合结构,在聚合代价体的同时自适应融合代价体和密集多尺度特征,从而使后续的解码层在多尺度上下文信息的引导下解码出更加精确和高分辨率的几何信息;最后,将全局优化后的高分辨率代价体送入视差回归模块以获得视差图。实验结果表明:所提算法在KITTI 2015和KITTI 2012数据集上的误匹配率分别降至1.76%和1.24%,SceneFlow数据集上的端点误差降至0.56 px,与GWCNet、CPOP-Net等先进算法相比,所提算法在不适定区域有明显改善。